Many children with autism spectrum disorders (ASD) and specific language impairment (SLI) have difficulties learning words, which negatively impacts academic performance, functional independence, and quality of life. Despite the importance of vocabulary development, relatively little is known about the learning mechanisms that underlie these deficits. Furthermore, the overlap of language characteristics in children with ASD and SLI is not well understood. An understanding of language-learning mechanisms in children with ASD and SLI has important theoretical implications upon which effective language therapies rely. Although statistical learning (i.e., sensitivity to linguistic regularities) is accpted as a robust language-learning mechanism in multiple domains of language in typical development, very little is known about statistical learning in children with ASD and SLI. This study will address this gap in knowledge by examining the impact of statistical learning in word learning. Specifically, this study will assess how children link knowledge of native language sound patterns to meaning. The contribution of statistical learning in word learning in children with atypical language development will be assessed through three studies. Specific Aim 1 will investigate prerequisite skills in word learning. Study 1a will assess the use of statistical learnng through a segmenting task (i.e., segmenting sound units from a continuous speech stream with the use of implicit transitional probability cues) without word learning. Study 1b will test associative word learning of novel labels paired with novel objects without segmentation. Because children with ASD and SLI have difficulties learning words and have difficulties consolidating linguistic information in word learning, it is important to test component skills in word learning. Once prerequisite skills have been assessed, Study 2, addressing specific aim 2, will assess whether children with ASD and SLI are able to map meaning onto newly segmented words. A significant relationship between prerequisite skills and specific aim 2 is anticipated; however, comparison among the different groups will allow for identification of potentially different abilities in word learning skills in the two populations with atypical language development. Both implicit and explicit teaching and testing methods will be used to capture subtle real-time learning information while placing minimal behavioral demands on the children. Within- and between- group analyses will assess learning in children with ASD, SLI, and typical development. New statistical methods for analyzing eye-gaze data will be employed to address potential analytical difficulties given the heterogeneity and possible lack of within-subject stability in children with ASD and SLI. Additionally, because of known social deficits in children with ASD, the studies will use non-social tasks. Findings from this study will contribute important information to our theoretical understanding of language acquisition in children with atypical language development.